1. Field of the Invention
This invention concerns a novel highly sensitive method for detection of the presence of total coliform and E. coli in water samples, food products, and seafood. Particularly, this invention concerns the method which employs new type of media having composition which allows an unequivocal detection of very small number of E. coli. The new media promote the growth of chlorine or food processing injured E. coli for their easier detection. This invention also concerns the low pH, trimethylamine-N-oxide containing media and the method for sole detection of E. coli in water, in food products and in seafood.
2. Background and Related Disclosures
Escherichia coli (E. coli) is a type species of Escherichia genus of Enterobacteriaceae family. E. coli is a gram-negative facultatively anaerobic bacteria normally inhabiting a gastrointestinal tract which bacteria is found in human and animal feces. E. coli is a frequent cause of infections of the urogenital tract and diarrhea.
Since E. coli is normally found in feces, its presence in the water or food stuff has been long recognized and used as an indicator of possible fecal contamination of water and food. Therefore, much effort has been expended in devising and improving methods for detection, identification and quantitation of these bacteria.
The use of E. coli as a water contamination indicator is disadvantaged, as there is no simple and specific way for their detection. Instead, coliform bacteria, a broader subset of species Enterobacteriaceae, which includes certain environmental organisms as well as fecal E. coli, but excludes Salmonella, Shigella and Proteus are usually detected. Consequently, coliform bacteria became a measure of a drinking water contamination by fecal matter.
Thus, it would be advantageous to have available an accurate test for determination of the presence of E. coli.
While there are available tests for determination of E. coli, such tests are not very accurate and may have a false positive or negative false error up 30-40%.
Currently available tests for determination of presence of coliforms and in particular E. coli are based on characterization tests which place the unknown organism in a defined group. Most commonly used tests are enzymatic tests where the presence of an enzyme known to be involved in the bacteria metabolism is detected subjectively by observation of a change in the appearance, a change in the color, a change in the fluorescence or some other chemical indication. Methods in Microbiology, 19:105 (1987).
Most of the existing tests for coliform and for E. coli are based on the last century findings described in Forsch. Med, 3:515 (1885), that Escherichia ferments glucose and lactose and produces gas or acid as a result of such fermentation. Most of the currently used methods utilize enzymes involved in the metabolism of lactose, such as, for example, the coliform bacterial enzyme .beta.-D-galactosidase which catalyzes the first step in the conversion of lactose to acid and gas.
Attempts to detect the presence of E. coli and differentiate it from other coliform early in the 20th century led to a classical fecal coliform (thermotolerant coliform) test involving incubation of bacterial cultures at 44.5.degree. C. At this temperature, most environmental coliforms will not grow but E. coli will grow and produce gas or acid from lactose.
Two primary disadvantages are associated with the fecal coliform technique. First, the procedure is lengthy. Second, due to rather high temperatures of above 44.degree. C. which must be used for bacterial incubation, the procedure is often lethal for stressed or injured bacteria. Since most of the drinking water is treated with chlorine which inevitably stresses and injures the E. coli cell membrane, under the normal circumstances the use of temperature over 44.degree. C. is not feasible.
To avoid this problem, it is customary to inoculate fecal coliform tubes with fully grown cultures from a presumptive coliform test. Such procedure takes two days. Alternatively, fecal coliform tubes may be incubated for a short while at 35.degree. C. followed by 44.5.degree. C. incubation. However, the gain in the shortening the test time is outweighed by additional manipulation involving timing of two incubations and additional handling. Second, as described in Appl. Environ. Microb., 55:335 (1989), the fecal coliform procedure has a limited specificity because it recognizes certain percentage of coliforms which are not E. coli as undesirable and harmful E. coli and thus results in false positive identification of E. coli where there is none.
In the clinics, many attempts to develop better tests for screening of fully grown pure colonies of pathogenic members of the family Enterobacteriaceae were made. In 1984, J. Clin. Microbiol., 20:136, described a single tube, multiple test medium for identification of fully grown pure cultures of Enterobacteriaceae from enteric and other clinical specimens. The medium used in that test allows detection of motility, .beta.-galactosidase, phenylalanine deaminase activity and hydrogen sulfide and indole production. However, the test can be done only on fully grown pure colonies and is thus not useful for water samples.
Lately, the presence of .beta.-glucuronidase (GUR) became a recognized indicator of the presence of E. coli. In 1976, Acta Pathol. Microbiol. Scand. Sect. B, 84:245 (1976) described findings showing that 97% of clinical isolates of E. coli produce .beta.-D-glucuronidase (GUR), whereas most other coliform bacteria do not. The success of GUR on clinical samples motivated scientists to apply it for E. coli detection in food and water.
One very serious flaw of the .beta.-glucuronidase procedure is that it presumes that all E. coli are able to produce .beta.-D-glucuronidase. Based on this presumption, a number of techniques, including the Rapid Identification Method (RIM) and Rapid Detect E. coli (RDA) for E. coli were developed for fully grown pure cultures of E. coli. The RIM technique, referenced below, involves simultaneous measurements of .beta.-glucuronidase and .beta.-galactosidase, followed by separate determination of the presence of indole. However, the technique does not allow the detection of weak or chlorine injured and MUG-negative E. coli and thus could result in approximately 30-40% false negative detection error if it were used to identify fecal E. coli.
A test for identifying fully grown pure cultures of E. coli, described in J. Clin. Mirobiol., 18:1287 (1983), consists of adding O-nitrophenyl-.beta.-D-galactopyranoside (ONPG) in Sorensen phosphate buffer having pH 7.5, to a fully grown test isolate and incubating the mixture for 1 hour at 35.degree. C. The presence or absence of E. coli is determined by presence or absence of a yellow color which indicates the presence of .beta.-galactosidase. If there is no color change in the original colorless solution, results are read as negative, and the sample is presumed not to contain coliform bacteria such as E. coli.
Two subsequently developed RIM and RDE tests are based on determination of .beta.-glucuronidase. While both these tests are recognized for detection of E. coli, they require fully grown colonies and are quite unsuitable for determination of small numbers of E. coli. Moreover, they are quite unable to detect E. coli which are injured by the chlorination or by the food processing. Both RIM and RDE tests are useful for determination of massive fecal E. coli contamination seen in hospitals but quite unsuitable for detection of small numbers of weak or injured E. coli typically occurring in drinking water treated with chlorine or following the food processing. Moreover, both tests are technically demanding and complicated.
The RIM system consists of reagents-impregnated cotton swabs on wooden sticks with which fully grown bacterial colonies are touched and the swab containing the bacterial inoculum is placed in specific RIM buffer containing ONPG, and incubated at 35.degree. C. The positive ONPG reaction, noted as the development of the yellow color, is followed by the determination of the presence of .beta.-glucuronidase activity by detecting the fluorescence upon irradiation with 366 nm UV light. If the .beta.-glucuronidase test is positive, Kovacs reagent is added for determination of indole presence indicated by red color.
The RDE test is a somewhat improved system for determination of .beta.-glucuronidase which utilizes a paper disk impregnated with substrates. In RDE, fully grown bacterial colonies are inoculated into distilled water to yield a bacterial suspension. Then, a disk containing both .beta.-glucuronidase and ONPG substrates are added to the tube which is incubated at 35.degree. C. Development of yellow color indicates the presence of .beta.-galactosidase, which hydrolyzes the synthetic substrate ONPG. Hydrolysis of ONPG indicates the presence of .beta.-galactosidase and thus the presence of lactose-fermenting organisms like Escherichia A-D group, Citrobacter, Klebsiella, and Enterobacter. When the .beta.-galactoside test is positive, the presence of .beta.-glucuronidase is determined, as in the RIM test, by the examination of the tube with a 366 nm UV light source. If there is a fluorescence, .beta.-glucuronidase is present. If the fluorescence is present, an indole indicator is then added as a third test to the sample to test for indole production. A red color indicates the presence of indole. Only when all three tests, i.e., ONPG, .beta.-glucuronidase and indole are positive, it is concluded that an organism is presumably E. coli. If any of the three tests is negative, the test is interpreted as denoting an organism having a low probability of being E. coli. In such an event, alternative procedures for confirmatory identification of clinical isolates are recommended and performed. J. Clin. Microbiol., 24:368 (1986).
Both RIM and RDE tests require purified, fully grown colonies and three independent indicators to be positive to determine the presence of E. coli. Consequently, such determination is still possible only in massively contaminated samples when E. coli is not injured by treatment or processing.
According to some earlier data published in Acta Pathol. Microbiol. Scand. Sect. B, 84:245 (1976), it was previously believed that .beta.-glucuronidase is present in about 97% of all clinical isolates of E. coli. The observation that about 97% of all clinical E. coli produce .beta.-glucuronidase led to the assumption that 97% of all E. coli from any source produce the enzyme. Recently, however, it has been found that the presumption of 97% glucuronidase-positive E. coli was false. New findings show that only about 66-70% of all E. coli from fecal samples are .beta.-glucuronidase-positive and about 30-34% of E. coli are .beta.-glucuronidase-negative.
These findings have recently been published in Appl. Environ. Microbiol., 55:335 (1989) showing that first, about 30-34% of all fecal isolates of E. coli are .beta.-D-glucuronidase-negative when measured with lauryl sulfate tryptose broth containing MUG, and second, among those .beta.-glucuronidase-positive E. coli, there is certain number of .beta.-glucuronidase-positive bacteria which are temperature dependent for .beta.-glucuronidase production. In these temperature dependent E. coli, such .beta.-glucuronidase production is rather weak at 37.degree. C. but strongly positive at 44.5.degree. C. Moreover, some fecal samples seem to contain only .beta.-glucuronidase-negative E. coli. Since the samples analyzed in the above study came from the fecal samples of healthy subjects, it is clear that the fecal contamination of water by E. coli may not be detected in around 30-35% of the time when currently available tests described above are used.
The previously accepted 97% glucuronidase-positive E. coli premise led consequently to a development of so called MMO-MUG (Minimal Medium ONPG-methylumbelliferyl-.beta.-D-glucuronidase) test currently approved by EPA as one of the analytical methods to enumerate total coliform.
The MMO-MUG test uses a medium containing a combination of ONPG and a fluorogenic compound MUG as a .beta.-glucuronidase substrate. The test is based on introducing a tested sample into the MMO-MUG medium and on observation of the production of .beta.-galactoside from ONPG, as evidenced by the formation of a yellow color in 24 hours, followed by development of a fluorescence due to metabolism of 4-methylumbelliferyl-.beta.-D-glucuronide, a substrate for enzyme .beta.-glucuronidase. The MMO-MUG test utilizes a specific multi-component medium consisting of ammonium sulfate, manganese sulfate, zinc sulfate, magnesium sulfate, sodium chloride, calcium chloride, potassium dihydrogen phosphate, sodium sulfate, anhydrous sodium phosphate, hydrogen sulphide, amphotericin B, ONPG, MUG, and Solanium. Several disadvantages have been recognized and connected with the use of the MMO-MUG test. First, the medium is expensive and currently may be obtained only from one source. It requires a presence of a proprietary dispersing agent Solanium not readily available. Second, the test is approved for estimating of total coliform bacteria as it is recognized that it is not specific enough to distinguish E. coli from other coliforms. Third, it requires the incubation at 35.degree. C. for at least 24 hours which incubation does not allow an addition of indole for specific determination of E. coli since it does not contain tryptophan, the substrate for indole production and since the 35.degree. C. temperature would allow a reaction of temperature sensitive indole positive non-E. coli bacteria if tryptophan were present. At temperatures above 44.degree. C., such reaction does not occur. Furthermore, the test has admittedly a false negative rate of least 13%, and more like 30-40%, since it does not detectMUG-negative E. coli, which error, for lack of better test, was found to be acceptable by the EPA.
In view of the 30-33% E. coli being .beta.-glucuronidase negative, the MMO-MUG test gives false negative results in about 30-34%. Such high false-negative rate of E. coli detection has serious implications for public health. As described in the Ann. Meeting Amer. Soc. Microbiol., Abstracts (1990), when E. coli contaminated samples of water from Southern California were examined, about 40% of the samples containing E. coli were found to be negative in MMO-MUG test.
The high false negative rate has been confirmed in other studies. A collection of fecal E. coli from many parts of the world was examined with both lauryl tryptose (LT) and the MMO-MUG technique, and the 30% false-negative rate was found, as described in the Journal of Food Protection, 53:972 (1990).
Thus, it would be highly advantageous to have available a specific and accurate test which would not depend solely on the determination of MUG-positive E. coli.
It is the subject of the current invention to provide such a test. The current method for determination of the contamination by E. coli in water, food, seafood or other samples is based on simultaneous measurement of lactose utilization and indole production by E. coli. The method is very specific for E. coli but allows the determination of the total coliform, fecal coliform and E. coli, if desired. In one aspect, method is modified to detect the presence of E. coli by using only the indole test.
The indole-producing enzyme tryptophanase was one of the first enzyme tests used for identification of E. coli. In the 100 years since Kitasato described the indole test in Zeitschrift Hyg. Leipzig, 7:515 (1889), it has proven to be one of the most useful and definitive tests for E. coli. It is used, as described in classical IMViC test published in Identification of Enterobacteriaceae, 2nd Ed. Burgess Publ. (1962), for distinguishing E. coli from other coliforms.
The major disadvantages of the indole test is that it cannot be combined with tests for lactose utilization because glucose and lactose both inhibit the indole production.
The principle of the indole test is the ability of E. coli to form tryptophanase, the enzyme which metabolizes tryptophan into indole. Some genera of the Enterobacteriaceae have the ability to produce tryptophanase and accumulate indole at 44.5.degree. C.; other genera do not. For example, E. coli produces indole at elevated temperature, but the related fecal coliform, Klebsiella pneumoniae, does not. Therefore, the accumulation of indole has proved to be useful and reliable way of distinguishing different members of the Enterobacteriaceae.
The formation of tryptophanase is inducible by tryptophan and its analogues. Since a number of species of Enterobacteriaceae has ability to produce tryptophanase, the presence of this enzyme results in the accumulation of indole as a catabolite of tryptophan and thus the presence of this enzyme is very useful for identification of Enterobacteriaceae bacteria.
The standard medium used for the identification of indole production, described in Standard Methods for Water and Wastewater Analysis requires incubation at 35.degree. C. for 24 hours in tryptone water, a carbohydrate-free medium, and testing for indole with Kovacs reagent (Zeit. Immunitaetforsch. Exp. Ther., 55:311 (1928)). This reaction is based on chromogenic reaction of indole with p-dimethylaminobenzaldehyde. In this test, free indole is removed by extraction to liquid media such as amyl alcohol or xylene to distinguish free indole from residual tryptophan, which also reacts with p-dimethylaminobenzaldehyde under suitable conditions.
An alternative chromogenic reaction for indole, described in J. Appl. Bacteriol., 63:329 (1987) is based on condensation of indole with glutaconic aldehyde to form red-violet polymethine dyes. While this is a reasonably rapid and sensitive method for determination of indole, the method requires a preparation of tryptophan agar media, lengthy incubation up to 48 hours at 37.degree. C., and the reaction with 1-(4-pyridyl) pyridinium chloride, a not readily available reagent. The later is used to saturate Whatman chromatography paper and needs to be converted to glutaconic aldehyde with sodium hydroxide. After rubbing a loopful of bacteria from a colony, already fully grown on tryptophan agar, on the Whatman paper, the acidification with hydrochloric acid is necessary to develop color reaction. This procedure is lengthy, laborious, requires rare reagents, and thus is not at all suitable for routine multiple-sample use for water testing.
A microtest, based on a similar idea, which is useful for rapid identification of Enterobacteriaceae is described in Acta Path. Microbiol. Immunol. Scand. Sect. B, 92:239 (1984). The test utilizes filter disks impregnated with a substrate on which a fully grown bacteria isolate is placed. The disk is reacted with a specific reagent to observe the development of colors characteristic for Enterobacteriaceae. Like other clinical diagnostic tests, this one requires a fully grown pure culture of the bacterium to be tested. It is laborious and time-consuming to make such isolates from water samples.
The deficiencies and disadvantages connected with the methods, tests and techniques currently available, as described above, are overcome with the method of the current invention. The current invention is based on the simple observation that tryptophanase, the enzyme responsible for indole production, is strongly repressed in the presence of lactose (Abstract: Ann. Meet. Amer. Soc. Microbiol., Q12 (1990)) and that this repression may be avoided by simply omitting lactose from the medium and replacing it with a suitable synthetic .beta.-D-galactopyranoside substitute such as ONPG (in this application, we use the term "galactoside" and "galactopyranoside" interchangeably). As described above, the ONPG serves to indicate the general ability of a bacterium to utilize lactose. However, unlike lactose, it does not repress tryptophanase or indole production and indole production can be detected by development of the purple color after addition of a suitable reagent such as Kovacs or Ehrlich reagent.